Natural Language Processing for Session-Based Information Retrieval Interface on the Web

Current web search engines enforce users to make a complete query at once, usually based on a few keywords, and rarely support query refinement process. However, internet search can most naturally be modeled as a session-based process, in which users can refine and modify their previous queries to satisfy their information needs. In this paper, we implement the idea of session-based web IR using natural language and discourse processing. The system supports natural language queries and commands along with a search history so that users can refine their queries based on the previous search results. The proposed system automatically generates standard boolean queries that can be interfaced with many currently available standard web search engines. Currently, large amount of query and command corpus are being collected to tune the NLP and discourse processing in the system. 1 Introduction Locating useful information on the web 1 still requires much patience and efforts. Current web sea..